Machine Learning Engineer - Predictive AI

GCS Recruitment; United States; On-site; USD 130,000 - 180,000; Full time

Job summary

Comcast is seeking a Machine Learning Engineer to own and refine predictive AI models for construction operations. You will take responsibility for the existing production model, develop new features, and build end-to-end pipelines to retrain models in production.

You will work with Python, XGBoost, AWS, MLflow, and modern MLOps tooling to push models from development into production, while analyzing errors and iterating on model performance.

Qualifications

• Strong hands-on Machine Learning Engineering experience.
• Experience building and improving predictive models.
• Strong understanding of feature engineering and model evaluation.
• Experience with Python and data pipelines.
• Experience with XGBoost or similar boosting algorithms.
• Experience working with production ML models, retraining, monitoring, and experimentation.
• Experience with AWS and modern ML tooling.
• Ability to investigate model performance and explain inaccuracies.

Responsibilities

• Take ownership of an existing production ML model and improve its performance across challenging use cases.
• Build and maintain pipelines to regularly retrain models.
• Analyze cases where the current model is less accurate and explore feature engineering or specialized models.
• Develop predictive models around construction timelines, permitting, and project costs.
• Work with historical data to identify patterns that can improve predictions.
• Engineer features around geography, job characteristics, historical response times, seasonality, holidays, and other factors.
• Experiment with different modeling approaches, including XGBoost and other machine learning techniques.
• Track experiments and model performance using MLflow.
• Work closely with the team to take ML models from development into production.

Skills

Machine Learning Engineering

Feature engineering

Model evaluation

Tools

Python

XGBoost

AWS

MLflow

DVC

FastAPI

HashiCorp Nomad

Job description

Machine Learning Engineer - Predictive AI

We're looking for a Machine Learning Engineer to join a team building predictive AI solutions for Comcast's construction operations.

The team already has a production ML model that predicts how long a construction job will take, and the model is currently performing well across the majority of use cases. We're looking for someone who can take ownership of the existing model, improve its performance, and help build the next generation of predictive models.

What You'll Work On

• Take ownership of an existing production ML model and improve its performance across challenging use cases.
• Build and maintain pipelines to regularly retrain models.
• Analyze the cases where the current model is less accurate and determine whether additional feature engineering or specialized models can improve results.
• Develop predictive models around construction timelines, permitting, and project costs.
• Work with historical data to identify patterns that can improve predictions.
• Engineer features around geography, job characteristics, historical response times, seasonality, holidays, and other factors.
• Experiment with different modeling approaches, including XGBoost and other machine learning techniques.
• Track experiments and model performance using MLflow.
• Work closely with the team to take ML models from development into production.

What We're Looking For

• Strong hands-on Machine Learning Engineering experience.
• Experience building and improving predictive models.
• Strong understanding of feature engineering and model evaluation.
• Experience with Python and data pipelines.
• Experience with XGBoost or similar boosting algorithms.
• Experience working with production ML models.Understanding of model retraining, monitoring, and experimentation.
• Experience with AWS and modern ML tooling.
• Ability to investigate model performance and determine why certain predictions are inaccurate.

Technology

Python | XGBoost | AWS | MLflow | DVC | FastAPI | HashiCorp Nomad | Machine Learning | Predictive Modeling | Feature Engineering

If you enjoy taking real-world business problems, turning them into predictive models, and then improving those models in production, this could be a great opportunity.